arXiv subjects

Jörg Kliewer

Publications and source records attributed to Jörg Kliewer.

At least 19 recordsLinked to original sources

arXiv2026

Context-Aware Search and Retrieval Under Token Erasure

This paper introduces and analyzes a search and retrieval model for RAG-like systems under {token} erasures. We provide an information-theoretic analysis of remote document retrieval when query representations are only partially preserved. The query is represented using term-frequency-based features, and semantically adaptive redundancy is assigned according to feature importance. Retrieval is performed using TF-IDF-weighted similarity. We characterize the retrieval error probability by showing that the vector of similarity margins converges to a multivariate Gaussian distribution, yielding an explicit approximation and computable upper bounds. Numerical results support the analysis, while a separate data-driven evaluation using embedding-based retrieval on real-world data shows that the same importance-aware redundancy principles extend to modern retrieval pipelines. Overall, the results show that assigning higher redundancy to semantically important query features improves retrieval reliability.

cs.IR↗

arXiv2026

Fundamental Limits of Coded Polynomial Aggregation

Coded polynomial aggregation (CPA) enables the master to directly recover a weighted aggregation of polynomial evaluations without individually decoding each term, thereby reducing the number of required worker responses. In this paper, we extend CPA to straggler-aware distributed computing systems and introduce a straggler-aware CPA framework with pre-specified non-straggler patterns, where exact recovery is required only for a given collection of admissible non-straggler sets. Our main result shows that exact recovery of the desired aggregation is achievable with fewer worker responses than required by polynomial coded computing based on individual decoding, and that feasibility is fundamentally characterized by the intersection structure of the non-straggler patterns. In particular, we establish necessary and sufficient conditions for exact recovery in straggler-aware CPA and identify an intersection-size threshold that is sufficient to guarantee exact recovery. We further prove that this threshold becomes both necessary and sufficient when the number of admissible non-straggler sets is sufficiently large. We also provide an explicit construction of feasible CPA schemes whenever the intersection size exceeds the derived threshold. Finally, simulations reveal a sharp feasibility transition at the predicted threshold, providing empirical evidence that the bound is tight in practice.

cs.IT↗

arXiv2026

Straggler-Aware Coded Polynomial Aggregation

Coded polynomial aggregation (CPA) in distributed computing systems enables the master to directly recover a weighted aggregation of polynomial computations without individually decoding each term, thereby reducing the number of required worker responses. However, existing CPA schemes are restricted to an idealized setting in which the system cannot tolerate stragglers. In this paper, we extend CPA to straggler-aware distributed computing systems with a pre-specified non-straggler pattern, where exact recovery is required for a given collection of admissible non-straggler sets. Our main results show that exact recovery of the desired aggregation is achievable with fewer worker responses than that required by polynomial codes based on individual decoding, and that feasibility is characterized by the intersection structure of the non-straggler patterns. In particular, we establish necessary and sufficient conditions for exact recovery in straggler-aware CPA. We identify an intersection-size threshold that is sufficient to guarantee exact recovery. When the number of admissible non-straggler sets is sufficiently large, we further show that this threshold is necessary in a generic sense. We also provide an explicit construction of feasible CPA schemes whenever the intersection size exceeds the derived threshold. Finally, simulations verify our theoretical results by demonstrating a sharp feasibility transition at the predicted intersection threshold.

cs.IT↗

arXiv2025

Performance Bounds on Pliable Index Coding Using Absent Receivers

We characterise bounds on the optimal broadcast rate for a few classes of pliable-index-coding instances. Unlike the majority of currently solved instances, which belong to a special class where all receivers with a certain side-information cardinality are either present or absent, we consider more general instances without this constraint. We devise a novel algorithm that constructs a decoding chain by iteratively adding a message that can be decoded by a receiver whose side information is already in the chain. If the decoding chain cannot proceed due to the absence of a receiver with the required messages, we skip a message by adding it to the chain regardless. We prove that a lower bound on the optimal broadcast rate is a function of the number of skipped messages, across all possible decoding choices of the receivers and any realisation of the algorithm for each decoding choice. While this result is not computationally feasible in isolation, it serves as a basis for deriving explicit lower bounds on the broadcast rate for specific classes of pliable-index-coding instances. These lower bounds depend on the number of absent receivers or the pattern of their side-information sets. Specifically, we explicitly characterise the optimal broadcast rate for instances with up to and including four absent receivers with any side-information pattern, as well as instances where the side-information sets are nested in particular ways.

cs.IT↗

arXiv2025

Communication-Constrained Private Decentralized Online Personalized Mean Estimation

We consider the problem of communication-constrained collaborative personalized mean estimation under a privacy constraint in an environment of several agents continuously receiving data according to arbitrary unknown agent-specific distributions. A consensus-based algorithm is studied under the framework of differential privacy in order to protect each agent's data. We give a theoretical convergence analysis of the proposed consensus-based algorithm for any bounded unknown distributions on the agents' data, showing that collaboration provides faster convergence than a fully local approach where agents do not share data, under an oracle decision rule and under some restrictions on the privacy level and the agents' connectivity, which illustrates the benefit of private collaboration in an online setting under a communication restriction on the agents. The theoretical faster-than-local convergence guarantee is backed up by several numerical results.

cs.SI↗

arXiv2025

Differentially-Private Decentralized Learning in Heterogeneous Multicast Networks

We propose a power-controlled differentially private decentralized learning algorithm designed for a set of clients aiming to collaboratively train a common learning model. The network is characterized by a row-stochastic adjacency matrix, which reflects different channel gains between the clients. In our privacy-preserving approach, both the transmit power for model updates and the level of injected Gaussian noise are jointly controlled to satisfy a given privacy and energy budget. We show that our proposed algorithm achieves a convergence rate of O(log T), where T is the horizon bound in the regret function. Furthermore, our numerical results confirm that our proposed algorithm outperforms existing works.

cs.IT↗

arXiv2025

Minimax Data Sanitization with Distortion Constraint and Adversarial Inference

We study a privacy-preserving data-sharing setting where a privatizer transforms private data into a sanitized version observed by an authorized reconstructor and two unauthorized adversaries, each with access to side information correlated with the private data. The reconstructor is evaluated under a distortion function, while each adversary is evaluated using a separate loss function. The privatizer ensures the reconstructor distortion remains below a fixed threshold while maximizing the minimum loss across the two adversaries. This two-adversary setting models cases where individual users cannot reconstruct the data accurately, but their combined side information enables estimation within the distortion threshold. The privatizer maximizes individual loss while permitting accurate reconstruction only through collaboration. This echoes secret-sharing principles, but with lossy rather than perfect recovery. We frame this as a constrained data-driven minimax optimization problem and propose a data-driven training procedure that alternately updates the privatizer, reconstructor, and adversaries. We also analyze the Gaussian and binary cases as special scenarios where optimal solutions can be obtained. These theoretical optimal results are benchmarks for evaluating the proposed minimax training approach.

cs.IT↗

arXiv2025

Context-Aware Search and Retrieval Over Erasure Channels

This paper introduces and analyzes a search and retrieval model that adopts key semantic communication principles from retrieval-augmented generation. We specifically present an information-theoretic analysis of a remote document retrieval system operating over a symbol erasure channel. The proposed model encodes the feature vector of a query, derived from term-frequency weights of a language corpus by using a repetition code with an adaptive rate dependent on the contextual importance of the terms. At the decoder, we select between two documents based on the contextual closeness of the recovered query. By leveraging a jointly Gaussian approximation for both the true and reconstructed similarity scores, we derive an explicit expression for the retrieval error probability, i.e., the probability under which the less similar document is selected. Numerical simulations on synthetic and real-world data (Google NQ) confirm the validity of the analysis. They further demonstrate that assigning greater redundancy to critical features effectively reduces the error rate, highlighting the effectiveness of semantic-aware feature encoding in error-prone communication settings.

cs.IR↗

arXiv2024

Differentially-Private Collaborative Online Personalized Mean Estimation

We consider the problem of collaborative personalized mean estimation under a privacy constraint in an environment of several agents continuously receiving data according to arbitrary unknown agent-specific distributions. In particular, we provide a method based on hypothesis testing coupled with differential privacy and data variance estimation. Two privacy mechanisms and two data variance estimation schemes are proposed, and we provide a theoretical convergence analysis of the proposed algorithm for any bounded unknown distributions on the agents' data, showing that collaboration provides faster convergence than a fully local approach where agents do not share data. Moreover, we provide analytical performance curves for the case with an oracle class estimator, i.e., the class structure of the agents, where agents receiving data from distributions with the same mean are considered to be in the same class, is known. The theoretical faster-than-local convergence guarantee is backed up by extensive numerical results showing that for a considered scenario the proposed approach indeed converges much faster than a fully local approach, and performs comparably to ideal performance where all data is public. This illustrates the benefit of private collaboration in an online setting.

cs.LG↗

arXiv2024

Straggler-Resilient Differentially-Private Decentralized Learning

We consider the straggler problem in decentralized learning over a logical ring while preserving user data privacy. Especially, we extend the recently proposed framework of differential privacy (DP) amplification by decentralization by Cyffers and Bellet to include overall training latency--comprising both computation and communication latency. Analytical results on both the convergence speed and the DP level are derived for both a skipping scheme (which ignores the stragglers after a timeout) and a baseline scheme that waits for each node to finish before the training continues. A trade-off between overall training latency, accuracy, and privacy, parameterized by the timeout of the skipping scheme, is identified and empirically validated for logistic regression on a real-world dataset and for image classification using the MNIST and CIFAR-10 datasets.

cs.LG↗

arXiv2023

On the Capacity of Private Nonlinear Computation for Replicated Databases

We consider the problem of private computation (PC) in a distributed storage system. In such a setting a user wishes to compute a function of $f$ messages replicated across $n$ noncolluding databases, while revealing no information about the desired function to the databases. We provide an information-theoretically accurate achievable PC rate, which is the ratio of the smallest desired amount of information and the total amount of downloaded information, for the scenario of nonlinear computation. For a large message size the rate equals the PC capacity, i.e., the maximum achievable PC rate, when the candidate functions are the $f$ independent messages and one arbitrary nonlinear function of these. When the number of messages grows, the PC rate approaches an outer bound on the PC capacity. As a special case, we consider private monomial computation (PMC) and numerically compare the achievable PMC rate to the outer bound for a finite number of messages.

cs.IT↗

arXiv2022

Pliable Private Information Retrieval

We formulate a new variant of the private information retrieval (PIR) problem where the user is pliable, i.e., interested in any message from a desired subset of the available dataset, denoted as pliable private information retrieval (PPIR). We consider a setup where a dataset consisting of $f$ messages is replicated in $n$ noncolluding databases and classified into $Γ$ classes. For this setup, the user wishes to retrieve any $λ\geq 1$ messages from multiple desired classes, i.e., $η\geq 1$, while revealing no information about the identity of the desired classes to the databases. We term this problem multi-message PPIR (M-PPIR) and introduce the single-message PPIR (PPIR) problem as an elementary special case of M-PPIR. We first derive converse bounds on the M-PPIR rate, which is defined as the ratio of the desired amount of information and the total amount of downloaded information, followed by the corresponding achievable schemes. As a result, we show that the PPIR capacity, i.e., the maximum achievable PPIR rate, for $n$ noncolluding databases matches the capacity of PIR with $n$ databases and $Γ$ messages. Thus, enabling flexibility, i.e., pliability, where privacy is only guaranteed for classes, but not for messages as in classical PIR, allows to trade-off privacy versus download rate. A similar insight is shown to hold for the general case of M-PPIR.

cs.IT↗

arXiv2022

Optimal Rate-Distortion-Leakage Tradeoff for Single-Server Information Retrieval

Private information retrieval protocols guarantee that a user can privately and losslessly retrieve a single file from a database stored across multiple servers. In this work, we propose to simultaneously relax the conditions of perfect retrievability and privacy in order to obtain improved download rates when all files are stored uncoded on a single server. Information leakage is measured in terms of the average success probability for the server of correctly guessing the identity of the desired file. The main findings are: i) The derivation of the optimal tradeoff between download rate, distortion, and information leakage when the file size is infinite. Closed-form expressions of the optimal tradeoff for the special cases of "no-leakage" and "no-privacy" are also given. ii) A novel approach based on linear programming (LP) to construct schemes for a finite file size and an arbitrary number of files. The proposed LP approach can be leveraged to find provably optimal schemes with corresponding closed-form expressions for the rate-distortion-leakage tradeoff when the database contains at most four bits. Finally, for a database that contains 320 bits, we compare two construction methods based on the LP approach with a nonconstructive scheme downloading subsets of files using a finite-length lossy compressor based on random coding.

cs.IT↗

arXiv2021

Private Function Computation for Noncolluding Coded Databases

Private computation in a distributed storage system (DSS) is a generalization of the private information retrieval (PIR) problem. In such setting a user wishes to compute a function of $f$ messages stored in $n$ noncolluding coded databases, i.e., databases storing data encoded with an $[n,k]$ linear storage code, while revealing no information about the desired function to the databases. We consider the problem of private polynomial computation (PPC). In PPC, a user wishes to compute a multivariate polynomial of degree at most $g$ over $f$ variables (or messages) stored in multiple databases. First, we consider the private computation of polynomials of degree $g=1$, i.e., private linear computation (PLC) for coded databases. In PLC, a user wishes to compute a linear combination over the $f$ messages while keeping the coefficients of the desired linear combination hidden from the database. For a linearly encoded DSS, we present a capacity-achieving PLC scheme and show that the PLC capacity, which is the ratio of the desired amount of information and the total amount of downloaded information, matches the maximum distance separable coded capacity of PIR for a large class of linear storage codes. Then, we consider private computation of higher degree polynomials, i.e., $g>1$. For this setup, we construct two novel PPC schemes. In the first scheme, we consider Reed-Solomon coded databases with Lagrange encoding, which leverages ideas from recently proposed star-product PIR and Lagrange coded computation. The second scheme considers the special case of coded databases with systematic Lagrange encoding. Both schemes yield improved rates, while asymptotically, as $f\rightarrow \infty$, the systematic scheme gives a significantly better computation retrieval rate compared to all known schemes up to some storage code rate that depends on the maximum degree of the candidate polynomials.

cs.IT↗

arXiv2019

Improved Lower Bounds for Pliable Index Coding using Absent Receivers

This paper studies pliable index coding, in which a sender broadcasts information to multiple receivers through a shared broadcast medium, and the receivers each have some message a priori and want any message they do not have. An approach, based on receivers that are absent from the problem, was previously proposed to find lower bounds on the optimal broadcast rate. In this paper, we introduce new techniques to obtained better lower bounds, and derive the optimal broadcast rates for new classes of the problems, including all problems with up to four absent receivers.

cs.IT↗

arXiv2019

Optimal-Rate Characterisation for Pliable Index Coding using Absent Receivers

We characterise the optimal broadcast rate for a few classes of pliable-index-coding problems. This is achieved by devising new lower bounds that utilise the set of absent receivers to construct decoding chains with skipped messages. This work complements existing works by considering problems that are not complete-S, i.e., problems considered in this work do not require that all receivers with a certain side-information cardinality to be either present or absent from the problem. We show that for a certain class, the set of receivers is critical in the sense that adding any receiver strictly increases the broadcast rate.

cs.IT↗

arXiv2019

Private Polynomial Computation for Noncolluding Coded Databases

We consider private polynomial computation (PPC) over noncolluding coded databases. In such a setting a user wishes to compute a multivariate polynomial of degree at most $g$ over $f$ variables (or messages) stored in multiple databases while revealing no information about the desired polynomial to the databases. We construct two novel PPC schemes, where the first is a generalization of our previous work in private linear computation for coded databases. In this scheme we consider Reed-Solomon coded databases with Lagrange encoding, which leverages ideas from recently proposed star-product private information retrieval and Lagrange coded computation. The second scheme considers the special case of coded databases with systematic Lagrange encoding. Both schemes yield improved rates compared to the best known schemes from the literature for a small number of messages, while in the asymptotic case the rates match.

cs.IT↗

arXiv2018

Capacity of Private Linear Computation for Coded Databases

We consider the problem of private linear computation (PLC) in a distributed storage system. In PLC, a user wishes to compute a linear combination of $f$ messages stored in noncolluding databases while revealing no information about the coefficients of the desired linear combination to the databases. In extension of our previous work we employ linear codes to encode the information on the databases. We show that the PLC capacity, which is the ratio of the desired linear function size and the total amount of downloaded information, matches the maximum distance separable (MDS) coded capacity of private information retrieval for a large class of linear codes that includes MDS codes. In particular, the proposed converse is valid for any number of messages and linear combinations, and the capacity expression depends on the rank of the coefficient matrix obtained from all linear combinations.

cs.IT↗